iflow-mcp_bashirk-inbound-mcp 0.1.0

A production-grade lead generation MCP server using FastMCP, Crawl4AI and external enrichment services

Lead Generation Server Documentation

Table of Contents

1. Overview
2. Features
3. Architecture
4. Prerequisites
5. Installation
6. Configuration
7. Running the Server
8. API Documentation
9. Examples
10. Advanced Configuration
11. Troubleshooting
12. Contributing
13. License
14. Roadmap
15. Support

---

Overview

A production-grade lead generation system built on:

• MCP Python SDK for protocol-compliant AI services
• Crawl4AI for intelligent web crawling
• AsyncIO for high-concurrency operations

Implements a full lead lifecycle from discovery to enrichment with:

• UUID-based lead tracking
• Multi-source data aggregation
• Smart caching strategies
• Enterprise-grade error handling

---

Features

Feature	Tech Stack	Throughput
Lead Generation	Google CSE, Crawl4AI	120 req/min
Data Enrichment	Hunter.io, Clearbit [Hubspot Breeze]	80 req/min
LinkedIn Scraping	Playwright, Stealth Mode	40 req/min
Caching	aiocache, Redis	10K ops/sec
Monitoring	Prometheus, Custom Metrics	Real-time

---

Architecture

graph TD
    A[Client] --> B[MCP Server]
    B --> C{Lead Manager}
    C --> D[Google CSE]
    C --> E[Crawl4AI]
    C --> F[Hunter.io]
    C --> G[Clearbit]
    C --> H[LinkedIn Scraper]
    C --> I[(Redis Cache)]
    C --> J[Lead Store]

---

Prerequisites

• Python 3.10+
• API Keys:

  export HUNTER_API_KEY="your_key"
  export CLEARBIT_API_KEY="your_key"
  export GOOGLE_CSE_ID="your_id"
  export GOOGLE_API_KEY="your_key"
  

• LinkedIn Session Cookie (for scraping)
• 4GB+ RAM (8GB recommended for heavy scraping)

---

Installation

Production Setup

# Create virtual environment
python -m venv .venv && source .venv/bin/activate

# Install with production dependencies
pip install mcp crawl4ai[all] aiocache aiohttp uvloop

# Set up browser dependencies
python -m playwright install chromium

Docker Deployment

FROM python:3.10-slim

RUN apt-get update && apt-get install -y \
    gcc \
    libpython3-dev \
    chromium \
    && rm -rf /var/lib/apt/lists/*

COPY . /app
WORKDIR /app

RUN pip install --no-cache-dir -r requirements.txt
CMD ["python", "-m", "mcp", "run", "lead_server.py"]

---

Configuration

config.yaml

services:
  hunter:
    api_key: ${HUNTER_API_KEY}
    rate_limit: 50/60s
    
  clearbit:
    api_key: ${CLEARBIT_API_KEY}
    cache_ttl: 86400

scraping:
  stealth_mode: true
  headless: true
  timeout: 30
  max_retries: 3

cache:
  backend: redis://localhost:6379/0
  default_ttl: 3600

---

Running the Server

Development Mode

mcp dev lead_server.py --reload --port 8080

Production

gunicorn -w 4 -k uvicorn.workers.UvicornWorker lead_server:app

Docker

docker build -t lead-server .
docker run -p 8080:8080 -e HUNTER_API_KEY=your_key lead-server

---

API Documentation

1. Generate Lead

POST /tools/lead_generation
Content-Type: application/json

{
  "search_terms": "OpenAI"
}

Response:

{
  "lead_id": "550e8400-e29b-41d4-a716-446655440000",
  "status": "pending",
  "estimated_time": 15
}

2. Enrich Lead

POST /tools/data_enrichment
Content-Type: application/json

{
  "lead_id": "550e8400-e29b-41d4-a716-446655440000"
}

3. Monitor Leads

GET /tools/lead_maintenance

---

Examples

Python Client

from mcp.client import Client

async with Client() as client:
    # Generate lead
    lead = await client.call_tool(
        "lead_generation",
        {"search_terms": "Anthropic"}
    )
    
    # Enrich with all services
    enriched = await client.call_tool(
        "data_enrichment",
        {"lead_id": lead['lead_id']}
    )
    
    # Get full lead data
    status = await client.call_tool(
        "lead_status",
        {"lead_id": lead['lead_id']}
    )

cURL

# Generate lead
curl -X POST http://localhost:8080/tools/lead_generation \
  -H "Content-Type: application/json" \
  -d '{"search_terms": "Cohere AI"}'

---

Advanced Configuration

Caching Strategies

from aiocache import Cache

# Configure Redis cluster
Cache.from_url(
    "redis://cluster-node1:6379/0",
    timeout=10,
    retry=True,
    retry_timeout=2
)

Rate Limiting

from mcp.server.middleware import RateLimiter

mcp.add_middleware(
    RateLimiter(
        rules={
            "lead_generation": "100/1m",
            "data_enrichment": "50/1m"
        }
    )
)

---

Troubleshooting

Error	Solution
403 Forbidden from Google	Rotate IPs or use official CSE API
429 Too Many Requests	Implement exponential backoff
Playwright Timeout	Increase scraping.timeout in config
Cache Miss	Verify Redis connection and TTL settings

---

Contributing

1. Fork the repository
2. Create feature branch: git checkout -b feature/new-enrichment
3. Commit changes: git commit -am 'Add Clearbit alternative'
4. Push to branch: git push origin feature/new-enrichment
5. Submit pull request

---

License

Apache 2.0 - See LICENSE for details.

---

Roadmap

• Q2 2025: AI-powered lead scoring
• Q3 2025: Distributed crawling cluster support

---

Support

For enterprise support and custom integrations:
📧 Email: [e-mail]
🐦 Twitter: @KobotAIco

---

# Run benchmark tests
pytest tests/ --benchmark-json=results.json